pivot_wider(): wrong names when names from numeric column + names_prefix

[mgirlich]

Doesn't work correctly for a numeric column

library(dplyr, warn.conflicts = FALSE)
library(dtplyr)
library(tidyr)

tibble(
  id = 1,
  name = 1:2,
  value = 11:12
) %>% 
  lazy_dt() %>% 
  pivot_wider(names_prefix = "col")
#> Source: local data table [1 x 3]
#> Call:   setnames(dcast(`_DT1`, formula = id ~ name, value.var = "value"), 
#>     old = 1:2, new = c("col1", "col2"))
#> 
#>    col1  col2   `2`
#>   <dbl> <int> <int>
#> 1     1    11    12
#> 
#> # Use as.data.table()/as.data.frame()/as_tibble() to access results

^{Created on 2021-03-04 by the reprex package (v1.0.0)}

but works when cast to character before

library(dplyr, warn.conflicts = FALSE)
library(dtplyr)
library(tidyr)

tibble(
  id = 1,
  name = as.character(1:2),
  value = 11:12
) %>% 
  lazy_dt() %>% 
  pivot_wider(names_prefix = "col")
#> Source: local data table [1 x 3]
#> Call:   setnames(dcast(`_DT1`, formula = id ~ name, value.var = "value"), 
#>     old = c("1", "2"), new = c("col1", "col2"))
#> 
#>      id  col1  col2
#>   <dbl> <int> <int>
#> 1     1    11    12
#> 
#> # Use as.data.table()/as.data.frame()/as_tibble() to access results

^{Created on 2021-03-04 by the reprex package (v1.0.0)}